AGN fueling : the observational point of view

Observations at multiple wavelengths are reviewed to search for evidence for fueling mechanisms in galaxies, both for nuclear star-bursts and AGN activity. Although it is undisputed that dynamical perturbations such as bars or tidal interactions accumulate gas in the central regions and sometimes trigger nuclear starbursts, the evidence remains scarce that these are necessary to fuel AGNs. Interpretations in terms of timescales , feedback , and black hole evolution are discussed. It is suggested that the AGN phase corresponds to the early-type phases of galaxies. 1. Fueling mechanisms The essential issue to fuel gas into nuclei is to find efficient processes to transfer angular momentum. Since viscous torques are not efficient except very close to the accretion disk itself, the principal actors here are gravity torques due to non-axisymmetric structures: 1. Bars, spirals, m=1 waves 2. Galaxy interactions and mergers These mechanisms are indeed observed when the AGN activity is very strong and accretion rate exceptional (> 10 M ⊙ /yr). For milder activities, it is difficult to see clear correlation with bars or companions. Nuclear starbursts and AGN have similar fueling requirements: they compete for gas supply and occur often simultaneously. The nuclear activity is however much more efficient to radiate energy and the fraction of energy contributed by the AGN (with respect to starbursts) increases with total luminosity. 2. Correlation with bars 2.1. Primary bars Many studies have addressed the role of bars in nuclear activity (Simkin et al 1980, Dahari 1984, Moles et al 1995). Weak correlations are found in general, and it is often difficult to discriminate the role of biases due to morphological types, environment, etc.. Bars are more easily detected in the old stellar component, and in near infrared surveys, free of dust extinction: many more barred galaxies are discovered then, but Seyferts are not privileged (McLeod & Rieke 1995, Mulchaey & Regan 1997). The bottom line of most surveys is that there is no difference of bar frequency between AGN and non-active galaxies (see fig. 1).